The linked SEER-Medicare data have rapidly become an important resource for measuring patterns and outcomes of cancer care. The advantage of using Medicare claims in addition to registry data is that they allow researchers to track care longitudinally and across sites. Investigators have begun to use Medicare claims to characterize particular chemotherapy drugs. The need to measure quality of care and increasing obstacles to medical record review indicate that SEER-Medicare will become an increasingly important resource for evaluating US cancer care. However, to date there have been few studies validating the use of SEER-Medicare to measure particular aspects of cancer treatment. Within the context of an IRB approved protocol, we propose Specific Aim 1: To validate Medicare claims as a resource for ascertaining use of: 1) intravenous chemotherapy treatment: 2) particular chemotherapeutic agents; and: 3) the duration and frequency of these treatments. We will validate the use of Medicare, claims for ascertaining chemotherapy use by comparing claims to "gold standard" pharmacy records from outpatient hemotherapy practices affiliated with Memorial Sloan-Kettering Cancer Center and the New York Hospital Network. This validation study will inform the health services research community about the [unreadable] reliability of using Medicare claims as a resource to understand the use of intravenous chemotherapy. [unreadable] If we determine Medicare claims have adequate sensitivity for specification of chemotherapy, we will [unreadable] address Specific Aim 2: To characterize the rate at which physicians adopt the use of new chemotherapeutic agents and to explore possible determinants of this process. Using SEER-Medicare data, we will describe how physicians adopted the use of chemotherapy drugs in response to FDA approval. We will specifically examine drugs approved in the mid-1990s including gemcitabine, paclitaxel, irinotecan and vinorelbine. We will characterize the rate at which physicians adopted each drug for specific cancers and, will identify patient, oncologist and drug-related parameters associated with slow diffusion. This knowledge has the potential to lay the groundwork for development of an inexpensive and efficient system for monitoring and improving the quality of US cancer care. [unreadable] [unreadable]